Today's electron sources integrated in power microwave amplifier tubes use the thermoelectronic emission obtained by heating electron sources called thermionic cathodes, to temperatures in the vicinity of 1000° C. Because of the physical principle used, these cathodes are limited in terms of emitted electron current and of lifespan, and also include the drawback of taking a fairly long time, of the order of a minute, to obtain the stabilized emission of electrons when they are heated, or switched on.
To circumvent these limitations and improve the efficiency of electron tubes with thermionic emission source, for example, in the case of power traveling wave tubes (TWT), solutions using an emission of electrons with the aid of cold semiconductor sources have been studied to replace the thermionic emission. These types of emissions by cold sources exploit the internal emission of avalanche ionization type or the field emission of tunnel effect type to emit or extract electrons from the semiconductor material.
In a first solution for producing cold electron sources, the emission of electrons is obtained from a PN diode made of silicon or of gallium arsenide, forward biased, the P zone being placed on the surface which is covered by a layer of cesium oxide. The role of this layer of cesium is two-fold:                to create a depletion zone by the dipole induced on the surface by this oxide where the electrons gain energy in the electrical field prevailing therein,        to lower the output work of the material in order to facilitate the emission of the electrons in the vacuum.        
Cesium oxide is, however, chemically unstable and the diode has to be made to work in a powerful vacuum to increase its lifespan. Even in these conditions, the layer of oxide degrades too rapidly for the device to be able to be used in the tubes. Furthermore, the maximum energy that the electrons can acquire is limited to the curvature of the bands in the vicinity of the surface and is, at best, of the order of the band gap of the materials used (typically less than 2 eV). The energy acquired by the electrons on passing through this zone is therefore less than the electron affinity of these materials which is of the order of 4 eV. Most of the electrons cannot therefore acquire sufficient energy to be emitted into the vacuum and only a small fraction, the most energetic of the electron distribution, leaves the material, hence low emission efficiency.
In a variant of this first solution, a metal of low electron affinity (LaB6 for example) replaces the layer of cesium oxide. The structure produced is used in diode mode, the electrical contacts being taken on the N-doped part of the diode and on the metal of low electron affinity. However, the gain in emission obtained by lowering the electron affinity using the material placed on the surface is wiped out by the energy losses induced by the collisions of the hot electrons with the network of the metal passed through.
A second solution uses a PN diode made of silicon or of gallium arsenide that is reverse biased beyond its avalanche breakdown voltage, the N zone being placed on the surface. In this approach, the current is obtained by avalanche multiplication and only the electrons that have an energy greater than the electron affinity of the material are emitted into the vacuum.
Given the semiconductors used, such devices have a very low emission efficiency. To increase the emission of electrons, a layer of cesium oxide is also deposited on the emissive surface but, as in the first solution, the instability of this oxide limits the lifespan of these devices.
A third avenue for producing cold electron sources exploits the field emission. In this solution the electrons are extracted from the material by tunnel effect using an intense external electrical field generated by point effect, either from molybdenum cones as in the Spindt cathodes, or from carbon nanotubes. These two solutions have not however led to any applications, the Spindt cathodes undergoing an accelerated degradation under the effect of the ion bombardment generated by the intense electrical field prevailing at the summit of the cones, and the carbon nanotubes not emitting a sufficient current density (effective emitted current density of the order of 1 A/cm2).
A fourth solution uses an NPN GaN bipolar transistor or the contact of the collector layer placed on the surface is pierced so as to allow for the emission of the electrons into the vacuum. The forward-biased base-emitter junction allows for the input of electrons, the reverse-biased base-collector junction makes it possible to provide the electrons with the energy needed for their extraction from the semiconductor. The impossibility of obtaining a strong concentration of holes at ambient temperature for P-doped GaN results in a high base access resistance value. This is reflected in the appearance of a phenomenon of lateral depolarization of the base-collector junction leading to a concentration of the current at the periphery of the component. The effective emissive surface is thus greatly reduced and represents no more than a small fraction of the total surface area of the transistor which results in a low emission efficiency.
None of the solutions described previously have to date made it possible to produce an electron source which is both reliable and intense enough to compete with the thermionic cathodes used today in the power tubes.